There exists a problem of co-channel interference between ZigBee and both wireless fidelity (WiFi, Wireless Fidelity) and Bluetooth. In addition, when ZigBee works at a 2.4 GHz frequency band, penetration and sensitivity of data transmission are relatively poor, and data transmission distance is also unsatisfactory. A spread spectrum technology used by WiFi has a great influence on ZigBee. ZigBee cannot work in places where there are WiFi hot spots. In a wireless meter reading system, a mainstream technology is a proprietary wireless ad-hoc network technology designed for a low-frequency ISM (Industrial Scientific Medical) frequency band, where the frequency bands include 915 MHz, 868 MHz, 433 MHz, and a newly released frequency band of 470-510 MHz. Networking in a low frequency band proves advantageous in terms of transmission distance (in particular, transmission distance at frequency bands 433 MHz and 470-510 MHz is five times longer than transmission distance of ZigBee), and can avoid co-channel interference with WiFi and Bluetooth. However, even if a frequency band of 470-510 MHz is used for communication, when a node is in an application scenario in which meters are deployed densely and in a complex and changeable cell electromagnetic environment, a communication link is also prone to unpredictable interference in time and space, which poses a great challenge to reliability of data interaction. According to features of the wireless meter reading scenario, a beacon interval in one personal area network (PAN, Personal Area Network) is relatively long and a duty cycle of an active period is relatively small. Therefore, channel quality determined by a gateway according to channel scanning defined in the GB/T15629.15-2010 standard cannot keep up with the pace of change in a relatively complex indoor electromagnetic environment (the channel quality changes quickly within a small time range). In addition, there is no request to send/clear to send (RTS/CTS, Request To Send/Clear To Send) handshake mechanism in GB/T15629.15. Consequently, it is likely that a channel which is dynamically selected according to the existing protocol is also exposed to interference from other interference signals suddenly after a next superframe period starts, which reduces a one-time data collection success rate. Using an adaptive frequency-hopping technology in combination with retransmission is a direct and effective approach to solving a problem of burst communication interference.
In the prior art, an adaptive frequency-hopping technology is used in combination with retransmission in the following two manners to solve the problem of burst communication interference:
First manner: An improvement is made to a beacon frame defined in the IEEE802.15.4 standard. For a star topology, a conventional 802.15.4 superframe structure is designed as multiple small superframes separated by a group acknowledge frame (GACK, Group acknowledge frame). The GACK frame further indicates that a new guaranteed timeslot (GTS, Guaranteed Timeslot) for a subsequent extended contention free period (ECFP, Extended Contention Free Period), and the new GTS is allocated to a node that needs to perform retransmission due to a previous transmission failure and a node that applies for a resource in the GTS temporarily to process a burst data stream. In this way, overheads during data interactions are reduced by way of GACK and data interaction is accelerated. For a point to point topology, a management timeslot is defined at the beginning of a superframe to inform a whole network of a newly joined/leaved full functional device (FFD, Full Functional Device) on a whole network, and a beacon timeslot in the FFD is also defined, in order to conveniently acquire communication information of a neighboring FFD. Based on information in a beacon frame, a superframe period is adjustable, a channel used in each announcement cycle (Announcement Cycle) may be different, and each PAN also uses a different channel.
However, in the foregoing technical solutions, an average data interaction rate can be increased by way of GACK, only on condition that channel quality is continuously good. Once a link failure occurs, some nodes cannot receive a GACK frame, and information in GACK messages will increase. If channel quality remains poor, a length of a GACK message increases cumulatively. In addition, if a node is disconnected from the network and a prescribed channel or a spare channel is not specified to listen to a beacon, re-initiating a network access process causes power usage to multiply.
Second manner: A two-level adaptive frequency hopping solution oriented to a clustering wireless sensor is used. For a two-layer topology that combines a star topology and a mesh topology and is applied on wireless networks for industrial automation-process automation (WIA-PA, Wireless Networks for Industrial Automation-Process Automation), content of a beacon frame payload is extended and a superframe is improved, in order to implement two-level adaptive frequency hopping, where a phase adaptive frequency hopping (PAFH, Phase Adaptive Frequency Hopping) is used at a beacon period, a contention access period (CAP, Contention Access Period) phase, and a contention free period (CFP, Contention Free Period) of the active period, that is, a same channel is used at these three phases within a same superframe period, and a channel is switched in different superframe periods according to a channel condition; a timeslot adaptive frequency hopping (TAFH, Timeslot Adaptive Frequency Hopping) is used at an intra-cluster communication phase of an inactive period, that is, a communication channel is changed, according to a channel condition, at each timeslot at the intra-cluster communication phase.
However, in the foregoing technical solution, it takes too long to switch a used channel in each superframe period to implement communication, due to a series of procedures such as measurement, judgment, and delivery, and therefore channel switching cannot keep up with the pace of change in channel quality; when an intra-cluster node or a cluster head node uploads measured channel information to the cluster head node or an aggregation node, a large amount of bandwidth resource is occupied and each node on the network needs to measure state information of the used channel periodically, which occupies a large cache capacity of a node device, reduces a processing speed of a device, and increases power consumption of the device.